Other Links

Indexed By

    •

Volume 19 Issue 2 - April 2021

  • 1. Quality control and assurance using image processing and opencv

    Authors : Rajas Mateti, Ritikesh Sawaimul

    Pages : 13-19

    DOI : http://dx.doi.org/10.21172/1.192.02

    Keywords : Image ProcessingOpenCVQuality ControlQuality AssuranceKeras-Tensorflow

    Abstract :

    This paper focuses on how to implement an automated defect identification system for infinite uptime devices using computer vision. In other words, we worked on the problem of detecting scratches, dents, and minor irregularities on an IoT device. This IoT device which comes with several sensors like temperature sensor, vibration, sensor, etc. is mounted on a number of mechanical devices and collects the required data. Due to the relatively small size of our dataset, we manually collected and annotated images. Various methods are used for generating more and more images. Here we dealt with an Image data generator which includes zooming, rotations, changing some angles, translation, etc a particular picture. Furthermore, more images were generated using various photoshop software. Because of the surface texture and curvature of sides, detection of dents, scratches can be a challenging task for traditional computer vision methods. The recent advancements in computer vision and the addition of many traditional libraries make it technically feasible to detect the dents and scratches which may not be visible by a naked eye. The project was divided into 3 models. The 1st part aimed to recognize the QR code and extract important data like the date of manufacturing, ID, etc. The 2nd part focuses on detecting the dis-alignment parts of the device and the 3rd part which is the main module detects the scratches, dents, and minor irregularities. OpenCV was used for dealing with the 2nd part, and the algorithm for the 3rd part was designed using Keras-TensorFlow. Results showed that dents, scratches that are not visible with the naked eye were effectively detected with 90% accuracy. Furthermore, we focused on generating more and more datasets for increasing accuracy. Our Study delivers a rigid introduction to image processing along with Image Data Generation techniques and computer vision fundamentals.

    Citing this Journal Article :

    Rajas Mateti, Ritikesh Sawaimul, "Quality control and assurance using image processing and opencv", Volume 19 Issue 2 - April 2021, 13-19

    Click here to Submit Copyright Takedown Notice for this article.

    View Paper